AI Interview Coach

flowchart LR
    subgraph Client
      FE[Next.js Frontend]
    end
    subgraph Gateway
      GW[Spring Cloud Gateway]
    end
    subgraph Services
      US[User Service]
      IS[Interview Service]
      AI[AI Service]
      SS[Statistics Service]
    end
    subgraph AI_Infra["AI Infrastructure"]
      LC[LangChain4j]
      LLM[Claude API]
    end
    subgraph Data
      PG[(PostgreSQL)]
      RD[(Redis)]
      CR[(ChromaDB)]
    end
    FE --> GW
    GW --> US
    GW --> IS
    GW --> AI
    GW --> SS
    AI --> LC
    LC --> LLM
    LC --> CR
    US --> PG
    IS --> PG
    SS --> PG
    IS --> RD
    AI --> RD

erDiagram
    users ||--o{ interview_sessions : "진행"
    users ||--o{ user_statistics : "보유"
    users ||--o{ daily_activity : "기록"
    job_descriptions ||--o{ generated_questions : "생성"
    interview_sessions ||--o{ interview_qna : "포함"
    generated_questions ||--o{ interview_qna : "사용"

    users {
      bigint id PK
      varchar email UK
      varchar password
      varchar nickname
      varchar target_job
      datetime created_at
    }
    job_descriptions {
      bigint id PK
      bigint user_id FK
      text jd_text
      varchar company_name
      varchar position
      json extracted_keywords
      datetime created_at
    }
    generated_questions {
      bigint id PK
      bigint jd_id FK
      text question_text
      varchar category
      varchar difficulty
      text reference_answer
    }
    interview_sessions {
      bigint id PK
      bigint user_id FK
      bigint jd_id FK
      enum status
      int total_questions
      int answered_count
      double avg_score
      datetime started_at
      datetime completed_at
    }
    interview_qna {
      bigint id PK
      bigint session_id FK
      bigint question_id FK
      text user_answer
      text ai_feedback
      double score
      int order_num
      datetime answered_at
    }
    user_statistics {
      bigint id PK
      bigint user_id FK
      int total_sessions
      int total_questions_answered
      double avg_score
      double best_score
      varchar strongest_category
      varchar weakest_category
      datetime last_updated
    }
    daily_activity {
      bigint id PK
      bigint user_id FK
      date activity_date
      int sessions_count
      int questions_count
      double avg_score
    }

sequenceDiagram
    participant C as Client
    participant GW as Gateway
    participant IS as Interview Service
    participant AI as AI Service
    participant LLM as Claude API
    participant CR as ChromaDB
    participant PG as PostgreSQL
    C->>GW: POST /api/interviews/jd
    GW->>IS: JD 분석 요청
    IS->>AI: JD 텍스트 전달
    AI->>LLM: 키워드 추출 요청
    LLM-->>AI: 직무 키워드
    AI->>CR: 키워드 임베딩 유사도 검색
    CR-->>AI: 관련 Q&A 컨텍스트
    AI->>LLM: 컨텍스트 + 질문 생성 요청
    LLM-->>AI: 맞춤 질문 목록
    AI-->>IS: 생성된 질문
    IS->>PG: 질문 저장
    IS-->>GW: 질문 목록 응답
    GW-->>C: 200 OK

sequenceDiagram
    participant C as Client
    participant GW as Gateway
    participant IS as Interview Service
    participant PG as PostgreSQL
    participant RD as Redis
    C->>GW: POST /api/interviews/sessions
    GW->>IS: 세션 생성
    IS->>PG: 세션 저장
    IS->>RD: 세션 상태 캐싱
    IS-->>C: 세션 ID + 첫 질문
    loop 각 질문마다
      C->>GW: POST /api/interviews/sessions/{id}/answer
      GW->>IS: 답변 제출
      IS->>PG: QnA 저장
      IS->>RD: 진행 상태 업데이트
      IS-->>C: 다음 질문
    end
    C->>GW: POST /api/interviews/sessions/{id}/complete
    GW->>IS: 세션 완료
    IS->>PG: 최종 결과 저장
    IS-->>C: 세션 요약

sequenceDiagram
    participant C as Client
    participant GW as Gateway
    participant AI as AI Service
    participant LLM as Claude API
    participant CR as ChromaDB
    participant RD as Redis
    C->>GW: GET /api/ai/feedback/{qnaId} (SSE)
    GW->>AI: SSE 연결
    AI->>RD: 캐시 확인
    alt 캐시 히트
      AI-->>C: SSE: 캐싱된 피드백 전송
    else 캐시 미스
      AI->>CR: 답변 임베딩 유사도 검색
      CR-->>AI: 참고 컨텍스트
      AI->>LLM: 스트리밍 평가 요청
      loop 토큰 단위
        LLM-->>AI: 토큰
        AI-->>C: SSE: data 토큰
      end
      AI->>RD: 전체 피드백 캐싱
      AI-->>C: SSE: [DONE]
    end